A ESP32-Cam based project that uses a PIR sensor to trigger photo upload to a web server. The PIR motion triggers are also broadcast using MQTT. The sketch runs a local webserver making live video streaming possible. It also publishes the ambient temperature readings from a connected temperature sensor. Configuration and camera settings are maintained through MQTT, and are stored in local SPIFFS files so they can survive a restart.
This started as a project to detect movement at a garden gate, that would then sound a notification in the house and switch on a garden light when dark. Once I decided to use a ESP32-Cam, it evolved to include photo capture and video streaming. I then slapped on a temperature sensor to upload the outside temperature to Home Assistant. Node Red is used to upload the captured photo to a Telegram bot.
- C++ sketch for a ESP32-Cam board.
- PIR motion sensor (AM312) events published using MQTT.
- Motion detection triggers photo capture and photo upload to a (PHP) web server.
- Temperature sensor (DS18B20) readings are published using MQTT.
- Runs a local webserver to allow realtime video streaming using e.g. MotionEye or Home Assistant (or just a browser).
- Camera settings can be maintained using MQTT. Camera settings are used to initialize the camera after startup.
- App configuration is maintained using MQTT. Configuration settings are used to initialize the device after startup.
- Camera and App settings are stored in JSON 6 format files in SPIFFs, to make the settings persistant and survive restarts.
- The board status, including WiFi strength, SoC Core temperature, time since last restart, etc. is published using MQTT.
Remarks:
- Confirm the ESP32-Cam board output voltage is 3.3V. Apparently the output pin voltage of some boards are set to 5V. This can be changed with a very small zero ohm resistor/jumper close to the output volate pin (may differ per board manufacturer).
- I powered both the AM312 and DS18B20 from the ESP32-Cam output voltage pin (the DS18B20 can also run on 5V).
- I had issues getting the DS18B20 to work. The temperature reading was -127°C, which is an error value indicating something went wrong. It worked for me using 3.3V instead of 5V, and putting a 4K7 pull-up resistor across the data pin and 3.3V.
| ESP32-Cam Pin | Description |
|---|---|
| GPIO 2 | DS18B20 Data Pin |
| GPIO 13 | AM312 Data Pin |
| 3.3V Output Pin | DS18B20 (+) Pin AM312 (+) Pin |
| Gnd | DS18B20 (-) Pin AM312 (-) Pin Power Source (-) |
| 5V | Power Source (+) |
- The white wires are only necessary when flashing the board with a new sketch, or when debugging in order to see the output of debug statements in the IDE.
Note: Ensure your FTDI card pins match the layout in the diagram! My (fake?) card pin layout was the opposite, an exact mirror image of the most common cards. - After development, when the board is deployed, connect the red and black wires providing power from the FTDI board during programing, to the positive (+) and ground (-) of a suitable 5V external power source.
| Function | Description | Values | Default |
|---|---|---|---|
| set_quality() | Quality | 10 - 36 | 10 (high) |
| set_brightness() | Brightness | -2 to 2 | 0 |
| set_contrast() | Contrast | -2 to 2 | 0 |
| set_saturation() | Saturation | -2 to 2 | 0 |
| set_special_effect() | Special Effect | 0 – No Effect 1 – Negative 2 – Grayscale 3 – Red Tint 4 – Green Tint 5 – Blue Tint 6 – Sepia |
0 |
| set_whitebal() | Enable White Balance | 0 – disable 1 – enable |
1 |
| set_awb_gain() | Enable White Balance Gain | 0 – disable 1 – enable |
1 |
| set_wb_mode() | White Balance Mode | 0 – Auto 1 – Sunny 2 – Cloudy 3 – Office 4 – Home |
0 (Auto) |
| set_exposure_ctrl() | Enable Exposure Control | 0 – disable 1 – enable |
1 |
| set_aec2() | AEC (night mode?) | 0 – disable 1 – enable |
0 |
| set_ae_level() | Automatic Exposure Level | -2 to 2 | 0 |
| set_aec_value() | Automatic Exposure Correction | 0 to 1200 | 204 |
| set_gain_ctrl() | Enable Gain Control | 0 – disable 1 – enable |
1 |
| set_agc_gain() | Gain Level | 0 to 30 | 5 |
| set_gainceiling() | Gain Ceiling (2x - 128x) | 0 to 6 | 0 |
| set_bpc() | Black Pixel Correction | 0 – disable 1 – enable |
|
| set_wpc() | White Pixel Correction | 0 – disable 1 – enable |
1 |
| set_raw_gma() | Enable Raw GMA | 0 – disable 1 – enable |
1 |
| set_lenc() | Lens Correction | 0 – disable 1 – enable |
1 |
| set_hmirror() | Horizontal Mirror | 0 – disable 1 – enable |
0 |
| set_vflip() | Vertical Flip | 0 – disable 1 – enable |
0 |
| set_dcw() | DCW (Downsize EN) | 0 – disable 1 – enable |
1 |
| set_colorbar() | Show Colorbar | 0 – disable 1 – enable |
- |
| face_detect() | Face Detection | 0 – disable 1 – enable |
- |
| face_recognize() | Face Recognition | 0 – disable 1 – enable |
- |
| set_framesize() | Resolution (0-10) | 10 UXGA(1600x1200) 9 SXGA(1280x1024) 8 XGA(1024x768) 7 SVGA(800x600) 6 VGA(640x480) 5 CIF(400x296) 4 QVGA(320x240) 3 HQVGA(240x176) 0 QQVGA(160x120) |
6 VGA(640x480) |
Reference:
The ESP32-Cam interacts with the outside world using MQTT.
The following is supported:
- Publish
- Movement Detection (can be enabled/disabled)
- Status Updates (periodic or when manually requested)
- Configuration Settings (on configuration change, or when manually requested)
- Camera Settings (only some key settings. when manually requested)
- Subscribed
- Trigger to take photo
- Update Camera settings (see table above)
- Update ESP32-Cam Configuration settings
- Request Status, Configuration
- Restart ESP32-Cam
See the MQTT Readme for detail on the MQTT setup and topics.
In my setup the ESP32-Cam is integrated with Home Assistant (HA), where the movement detection is displayed and the temperature readings are captured. Capturing a photo can also be triggered from HA via MQTT.
See the Home Assistant Readme for more detail.
Some of the board automations are handled in Node Red.
See the Node Red Readme.
In my implementation the ESP32-Cam uploads captured photo's to a PHP server, from where Node Red will publish each newly captured photo to a Telegram bot.
See the PHP Readme